Key takeaways

• OpenAI’s compute margins rose to around 70% by October 2025, nearly doubling from the previous year.
• ChatGPT mobile app revenue surpassed $3 billion, with $2.5 billion generated in 2024 alone.
• Despite efficiency gains, OpenAI remains unprofitable as infrastructure costs and competition increase.

Efficiency improvements drive stronger margins

OpenAI has made notable gains in operating efficiency, reportedly doubling its compute margins to approximately 70% by October 2025. That figure reflects the share of revenue retained after covering the cost of deploying and running its AI models for paying users—a key metric of product-level profitability. A year earlier, those margins stood near 52%.

The jump, cited in a report by The Information, comes as OpenAI works to reduce reliance on expensive compute resources and refine backend model performance. Bloomberg sources noted that OpenAI now outpaces Anthropic in margins for paying users, despite Anthropic’s more streamlined server infrastructure.

Even with these improvements, the company remains in the red. OpenAI’s valuation, pegged at $500 billion in several private market assessments, has spurred concerns about financial sustainability. Internal discussions, according to reports, have described a “critical situation” as executives push to accelerate development and keep pace with competitors like Google and Anthropic. Most recently, OpenAI has been exploring new funding that could lift its valuation to as high as $830 billion.

Consumer spending on ChatGPT accelerates

ChatGPT’s mobile app has proven to be a key revenue driver. Since launch, users have spent more than $3 billion through the app, with over $2.5 billion of that occurring in 2024 alone, according to data from Appfigures. That growth represents a 408% year-over-year increase and places ChatGPT ahead of several social and streaming platforms in terms of time-to-revenue benchmarks.

The app’s rapid success points to strong consumer demand for AI-driven productivity and conversational tools. Analysts view the scale and speed of adoption as signs that AI services are becoming embedded in daily digital behavior.

AI adoption reshapes consumer routines

Emerging data highlights just how deeply tools like ChatGPT have penetrated global routines. Research by PYMNTS estimates up to 900 million people now interact with ChatGPT weekly, generating roughly 2.5 billion queries per day. By contrast, it took Google over a decade to reach comparable usage levels.

In the U.S., more than half the population reportedly uses AI chatbots for a range of tasks, including shopping, learning, and health-related inquiries. Around 30 million users interact with chatbots across at least 25 distinct activities tied to over 50 sectors in the connected economy.

Among those users, more than 80% rely on AI tools to assist with product discovery, planning, education, and wellness—suggesting a shift in how people access and consume digital services.

The bottom line

OpenAI’s progress in driving down costs and growing mobile revenues signals major strides in its commercial roadmap, but fundamental challenges remain. The company continues to rely heavily on Microsoft Azure for infrastructure, which could limit its ability to cut costs further. Broader economic factors—from high interest rates to regional energy constraints—also complicate its path to profitability.

Meanwhile, rivals including Apple, Meta, and Google are investing in proprietary AI chips, which could tilt the economics of AI infrastructure in their favor. Whether OpenAI can maintain its lead without vertically integrated solutions remains an open question, even as it cements its role as a dominant force in the AI landscape.